The job of a direct mail marketer is vital to many brands but has become increasingly challenged by colliding factors such as the still-uncertain economic future and constantly growing postage costs. These challenges represent more reasons mailers should be doing everything they can to target qualified prospects, manage campaign costs, and drive profitable results. Enter direct mail optimization.
Optimization harnesses predictive modeling to identify which consumers in your mail plan are most likely to take a desired action. Commonly optimizing for response, mailers can see tremendous lift in campaign KPIs and a reduction in cost as they only mail consumers who rank highly in the model. Despite these tremendous results, many marketers are often hesitant to invest in yet another step in the direct mail process. One excellent way to explore the benefits of optimization without risk is to back test first.
Back testing is a great way to understand how optimization can impact both costs and results of adding this additional layer. It is a process that Alliant has executed many times for direct mail partners to help highlight the benefits. To start, a brand does need to do a bit of legwork and provide historical data, costs, and KPIs. Be prepared to supply:
Two historical campaign files
One file should be exemplary of a campaign with strong results. This will serve as the as the seed data for the model. The second file should be another mail file for the same product and offer but a different campaign date. Both files need to include detailed response and order details.
A KPI that matters
As previously noted, this could be response, but it could also be 2+ orders. Or, if you’re looking to mail to expired customers, perhaps it’s reactivation with a new purchase.
Campaign costs
Providing detailed costs on the mail piece, postage, lists costs, and fulfillment enables a cost savings analysis to prove the ROI of optimization.
With these assets, a data science team will build a custom model. Using the custom model, they will score all the consumers mailed in the second historical campaign file. The results of this scoring will then be compared to the actual results of that campaign. Continuing with response as our KPI example, analysis will note how consumers actually responded, and who did not, against where they fell in the model ranks. The top ranks of the model should have predicted responders at a high rate, while the bottom ranks will include consumers with no response.
In a back test, you aim to prove that the model accurately identified the responders and the non-responders. Since you compare this to actual campaign results, it makes it easy to verify if the model is accurately predicting the consumer action in question. The costs analysis then applies the campaign investment to those bottom groups and illustrates cost savings had you decided not to ever mail them. In summary, the analysis will provide the metrics of total campaign response, total costs, and total revenue, for the scenario with optimization and without.
Yes, there is a bit of legwork required. However, not only will your optimization partner execute the analysis and present findings, but you will also receive a custom model – built, tested, and ready to use on your next campaign – at the end of this process. For those that mail with great frequency, or those who mail great quantities, making the investment in time and resources to appropriately explore an optimization back test can yield a tremendous return on investment. A back test may also be appropriate for brands that are already optimizing but looking to expand their optimization efforts with a new partner.
Interested in setting up your back test today? Reach out to the Alliant help desk or your Account Executive.
Submit a Comment